What problem does it solve? Writing effective prompts for Higgsfield AI video generation is difficult: generic prompts produce inconsistent results, identity drifts across shots, and wasted credits accumulate from unfocused iteration. This Skill provides a structured prompt-engineering framework so every generation starts from a well-formed prompt. ## Core Features & Use Cases - MCSLA Formula: Structures every prompt across five layers — Model, Camera, Subject, Look, Action — with genre-specific length and lead-with targets. - T2V and I2V Patterns: Provides distinct templates for text-to-video scene building and image-to-video animation, including the rule to describe only what moves in I2V. - Identity vs. Motion Separation: Splits prompts into Identity Blocks and Motion Blocks to keep Soul ID characters consistent across shots. - Diagnostics and Iteration: Includes the 6-Pass Diagnostic Sequence, single-variable iteration rule, and batch-and-select guidance for stochastic misses. - Use Case: A creator needs a cinematic product shot of a rotating earbud case. The Skill routes to the Product genre (30–50 words, subject-led), applies MCSLA structure, names a camera preset, and enforces the Cinema Studio 512-character cap. ## Quick Start Ask the agent to write a Higgsfield prompt for your scene, specifying the subject, action, camera move, and visual style you want.
